首页> 外文期刊>Acta acustica united with acustica >Assessing Speech Comprehension in Noise: Acoustic Quality of Speech for Near Real Communication Conditions
【24h】

Assessing Speech Comprehension in Noise: Acoustic Quality of Speech for Near Real Communication Conditions

机译:评估噪声中的语音理解:接近真实通信条件下的语音声学质量

获取原文
获取原文并翻译 | 示例
           

摘要

To assess the quality of speech communication, reference is usually made to speech intelligibility, but what does that mean for practical speech conditions? Even with high signal-to-noise ratios speech intelligibility alone falls short of what is required for a broader scope of differentiation. The pursued objective here is to improve the description of the communication situation which means paying more attention to the perception and comprehension processes with the main focus on the listener's efforts to cope with the situation or the stress and strain he experiences. In two experiments - on speech intelligibility and the assessment of the comprehension process listeners were presented sentences and monosyllables at six different speech levels (40 to 75 dB) and varying traffic noises (55 dB). For each sentence both speech intelligibility and answers to assessment questions concerning coping, concentration, annoyance, and subjectively rated intelligibility (on a 5-point scale) have been evaluated. The answers to the assessment questions always refer to three different sentences with the same experimental conditions. As a result, assessment parameters have shown to be more sensitive, especially in cases where speech intelligibility is nearly constant (for levels of high speech intelligibility: SI > 80%, SNR > 5 dB). The data indicates that not until a signal-to-noise ratio of beyond SNRA > 15 to 20 dB high intelligibility with lowest possible mental effort can be achieved. The noise level has a particular influence on the level of annoyance (always at the same SNR) whereas a simultaneously applied task only affects one's concentration. Based on the assessment of coping and concentration five quality levels for the comprehension process have been developed.
机译:为了评估语音交流的质量,通常会参考语音清晰度,但这对实际的语音条件意味着什么?即使在高信噪比的情况下,仅语音清晰度也不能满足更大范围的差异化要求。这里追求的目标是改善对交流情况的描述,这意味着更多地关注感知和理解过程,主要集中在倾听者为应对这种情况或他所经历的压力和压力而付出的努力。在两个实验中-关于语音清晰度和对理解过程的评估,听众在六个不同的语音水平(40至75 dB)和变化的交通噪音(55 dB)下呈现了句子和单音节。对于每个句子,都评估了语音清晰度和对应对,注意力,烦恼和主观评估清晰度(5分制)的评估问题的答案。评估问题的答案总是指具有相同实验条件的三个不同句子。结果,评估参数显示出更加敏感,尤其是在语音清晰度几乎恒定的情况下(对于高语音清晰度:SI> 80%,SNR> 5 dB)。数据表明,直到信噪比超过SNRA> 15至20 dB时,才能以尽可能低的精力完成高清晰度。噪声级别对烦扰程度(始终处于相同的SNR)有特殊的影响,而同时执行的任务只会影响一个人的注意力。基于应对和集中力的评估,已经为理解过程开发了五个质量等级。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号